Latency, Jitter, and Loss – How to Speak Clearly in the Cloud

The new age of telephony

As the Internet has evolved, so has its impact on our culture. First it was ­music and shopping, then movies and TV shows, today, it’s IP voice communication in the Cloud.  While cloud communication services can be significantly more economic to deliver and use than legacy voice services, one should not have to sacrifice quality to get those benefits.  It is important to know the details behind voice quality.

The central issue for cloud-based voice communication is that speaking is done in “real-time” and the Internet is a packet-switched network.  Packet switching means that voice samples are put into IP packets, addressed and sent on their way; not unlike one might do when mailing a series of letters with each of those letters containing part of a conversation.  Just like letters dropped in the mail, information doesn’t always follow the same path in packet-switched networks and packets of information (parts of the conversation) can get lost, delayed, or arrive out of sequence. In real-time communication, there simply isn’t temporal room for lost information, thus annoyances such as jittery words, clipping, and echoes are the result.

Meet your enemies: Latency, jitter, and packet loss

The big three enemies of poor voice quality are latency, jitter and packet loss.  Let’s define each of them.

Latency.  Latency measures the delay of a packet from point A to point B, or how long it takes for a packet to reach its destination – sound quality is negatively impacted by latency.  Once delays of over 250 milliseconds are reached, most callers report frustrations with their ability to communicate effectively.  Delay of 100ms or less is preferable with studies showing that when it drops below 100ms, it is no longer appreciably perceived by the listener.

Jitter.  Jitter is the variance in the delay of received packets (some arrive in 75 milliseconds, some in 200 milliseconds, some in 100 milliseconds etc..).  As you might imagine, this has a jarring effect on voice quality. Generally speaking, jitter can occur during times of heavy network traffic and/or slow network speeds. Jitter buffers are a popular fix. They gather packets in a buffer and reassemble them in their proper order and meter them out smoothly – leaving the listener none-the-wiser.

Packet loss. Packet loss can occur when the network becomes congested or when a portion of the path which a series of packets takes becomes congested.  When a network handles video, data, and voice, it can become overwhelmed if it is not engineered properly.  When this occurs, packets are thrown away and don’t arrive at their destination until the congestion calms down and traffic flows more easily.  If the communication is not “real-time”, this is not an issue because those missing packets can be resent.  But voice is real-time, and you don’t get a second chance to send a missing packet.  The result of those missing packets is that parts of words or parts of a conversation never arrive at the destination and so it sounds like chunks of the conversation are missing.

Bringing it all together

The good news is these problems are solvable if your cloud communications provider has a solution for delivery of high quality voice.   It demands a unique combination of network infrastructure, customer premise equipment, voice quality measurement, definitive trouble shooting and fault isolation tools combined with deep technical expertise. When you can control latency, packet loss, jitter, you can deliver a quality voice experience.

As an early pioneer in cloud communications and with over a decade of commercial service under our belt, nexVortex has developed a complete portfolio of managed and unmanaged SIP Trunking Services, Hosted Voice, and Hosted Contact Center services, all capable of meeting the most discerning customer demand.  Contact one of our experts today to experience how you can have the voice quality your business deserves.